Data prep

COVID-19 breakthrough cases among Harford County residents in Maryland

(Page updated at 2021-11-01 09:50:15)
Data from:
* COVID Link breakthrough cases data by MDH, as of 10.29.21
* COVID Link VOC/VOI data by MDH, as of 10.29.21
* The number of total new COVID-19 cases by MDH, as of 2021-11-01.
* Vaccination data from Immunet, as of 10292021.
* COVID19 hospitalization data by MDH, as of 10.29.21.
See Annex for detailed note about the data.

1. What is the level of breakthough cases?

1.1. Trend of breakthrough cases - absolute number

A total of 42154 COVID-19 breakthrough cases have been identified in Maryland. In Harford county, 2217 cases have been reported. Below figure shows the daily number of breakthrough cases over time.

(NOTE: the 7-day average is the mean during a week leading to the date.)

1.2. Trend of breakthrough cases - percentage

To date, among people who have been fully vaccinated in the county (n=150746), 1.47% have had a breakthrough case - assuming no multiple breakthrough cases per person.

Following figure shows cumulative percent of fully vaccinated people who have breakthrough cases. We expect the cumulative percent to:
- fluctuate initially, depending on the pace of numerator increase (number of breakthrough cases - see above figure) vs. pace of denominator increase (number of people fully vaccinated), and then
- level-off, based on expected vaccine effectiveness.

(Rate using person years forthcoming… )

(Note: percent = 100* ‘# cum cases on day X’ / ‘# cum fully vaccinated people on day X-14’)

1.3. Level of new cases

Note:
* Non breakthrough cases = total cases - breakthrough cases. This assumes date in total cases is based specimens collection date, which was used in breakthrough case data. This may not be correct, but the implication is minor as long as there is only minimal report delay.
* Fully vaccinated population on day X = population who received second/final dose 14 days ago (X-14).
* Not fully vaccinated population = total population - fully vaccinated population * See detailed numerators and denominators below.

1.4. Can we estimate vaccine effectiveness - to prevent cases?

Below figure shows expected percent of breakthrough cases based on vaccine efficacy and vaccine coverage, according to Chen and Orenstein (1996). But, this method requires routine (unbiased) surveillance system…

2. What are characteristics of breakthough cases?

2.1. Timing of breakthough cases since the second/final dose

In Harford county, 50% of breakthrough cases occurred within 150 days after receiving the second dose. 75% of breakthrough cases occurred within 188 days.

In Maryland, similarly, 50% of breakthrough cases occurred within 147 days after receiving the second dose. 75% of breakthrough cases occurred within 184 days.

But, by age group, the distribution is little different. See below.

2.1.A. Survival analysis

No ID to merge immunet data and breakthrough data. need to check all COVID link reports variables.

2.2. Other characteristics among breakthrough cases by period: prior to vs. during the last 6 weeks

COVID-19 breakthrough cases by group: prior to vs. during the last 6 weeks
Group Group.label In the last 6 weeks: Harford county, Number In the last 6 weeks: Harford county, % In the last 6 weeks: MD, Number In the last 6 weeks: MD, % Prior to the last 6 weeks: Harford county, Number Prior to the last 6 weeks: Harford county, % Prior to the last 6 weeks: MD, Number Prior to the last 6 weeks: MD, %
By Age
12-19 53 4 1325 5 55 7 778 5
20-29 140 10 3629 14 91 11 1640 12
30-39 257 19 4788 18 121 15 2449 17
40-49 278 20 4619 18 150 18 2326 16
50-59 231 17 4493 17 161 19 2615 18
60-69 210 15 3926 15 143 17 2347 16
70+ 220 16 3499 13 107 13 2091 15
11 and Under 19 0 11 0
By death
No 1373 99 26059 99 825 100 14183 99
Yes 16 1 239 1 3 0 74 1
By hospitalization
No 1388 100 26226 100 816 99 14038 98
Yes 1 0 72 0 12 1 219 2
By Having.Symptoms
missing 148 11 2427 9 370 45 5157 36
No 686 49 10441 40 33 4 999 7
Yes 555 40 13430 51 425 51 8101 57
By Variant.Lineage
Not tested 1188 85.53 22068 83.92 806 97.34 13982 98.07
VOC: Alpha (UK) 20 1.44 341 1.3
VOC: Beta (South Africa) 10 0.04
VOC: Delta (India) 114 8.21 2730 10.38 12 1.45 160 1.12
VOC: Gamma (Japan/Brazil) 19 0.07
VOI: Eta 7 0.03
VOI: Iota (US/NY) 5 0.36 49 0.19
Wild Type 11 0.79 145 0.55
NA 51 3.67 929 3.53 10 1.21 115 0.81
By Congregate.Housing
0 1351 97 25809 98 820 99 14004 98
1 38 3 489 2 8 1 253 2

2.3. Characteristics comparison: breatkthrough vs. non-breakthrough cases

Work in progress using Redcap/Prepmod data

2.4. Hospitaization: breakthrough vs. regular cases

NEDSS is unique in redcap data.
NEDSS is NOT unique in hospitalization data…

Work in progress. YJ to double check ID variables in COVIDlink reports…

4. Among Harford county residents who were hospitalized, vaccination status at the time of specimen collection - weekly trends

Below figures are weekly trends for Harford county, based on MDH’s “Weekly COVID Hospitalizations Line List”, received on 10.29.21.

4.1. Broken down only by vaccine status

4.2. Broken down by vaccine status AND age group

4.3. Broken down only by age group

4.4. Broken down only by detailed age group

###Annex Data source 1. Breakthrough cases: Reports, Cases w Vax Breakthrough (>=14 days) 4 and Cases w Vax Breakthrough (>=14 days) 3, from COVID Link (http://mdcovid.lightning.force.com/).

Breakthrough cases are identified, based on the following - according to Alex Goode at MDH:
“1. COVID-19 case is matched to the CRISP master patient index based on Firstname, Lastname, DOB, Gender, Address - pulls in something called EID - Enterprise Identifier
2. Immunet data is separately matched to the CRISP master patient index on Firstname, Lastname, DOB, Gender, Address
3. Immunet Data is added to the contact object where contact.EID = immunet.EID
4. We provide the Immunet Names and DOB to ensure the match is accurate, since there is a very small chance of mismatching, although this match is usually > 99% reliable and catches variations in name and address entry, and accommodates simple transpositions of dates of birth, etc in one of the sources - still allowing the records to match.”

Data source 2. New COVID-19 cases: from MDH (https://opendata.arcgis.com/datasets/0573e90adab5434f97b082590c503bc1_0.geojson)

Data source 3. Number of people fully vaccinated: Immunet data from MDH.

Data source 5. COVID-19 hospitalization data; updated data from MDH received on Fridays.

Data source 5. VOC/VOI cases: A “report”, Cases with VoC, from COVID Link (http://mdcovid.lightning.force.com/).

Also, see below for further information about testing for variants (source: Email communication with Monique Duwell, MDH, Chief, Center for Infectious Disease Surveillance and Outbreak Response; July 29, 2021)

"1. Can you tell us roughly what percent of COVID-19 cases are tested for variants in Maryland? Currently, and, if different, in the past as well?
Based on the sequencing results reported to our surveillance program, about 15% of COVID cases are being sequenced. This is likely an under-estimate of the true % of sequencing being performed statewide.

  1. Is the testing done randomly or more focused among certain patients (e.g., by age, by vaccination status, by severity/admission status)?
    We focus on sequencing patients who are hospitalized, deceased, fully vaccinated, live in high-risk settings, and who have traveled. However, we are also obtaining sequencing results on many patients who do not fall into these categories, since some labs sequence all patients, regardless of risk factors.

3.Do you know the general turnaround time for the variant test result: from specimen collection to data entry in the COVID link?
It varies greatly, largely due to the variability in amount of time it takes for labs to complete sequencing, and the amount of time it takes for them to upload the data to GISAID or report to us directly. I would say that about 3-4 weeks is the average turnaround time."